Re: [PATCH v2] drm/amdgpu: fix system hang issue during GPU reset

[Andrey Grodzovsky <Andrey.Grodzovsky@xxxxxxx>]

On 7/8/20 3:48 AM, Dennis Li wrote:
> During GPU reset, driver should hold on all external access to
> GPU, otherwise psp will randomly fail to do post, and then cause
> system hang.
>
> v2:
> 1. add rwlock for some ioctls, debugfs and file-close function.
> 2. change to use dqm->is_resetting and dqm_lock for protection in kfd
> driver.
> 3. remove try_lock and change adev->in_gpu_reset as atomic, to avoid
> re-enter GPU recovery for the same GPU hang.


Still adev->in_gpu_reset is prone to race, no ? I mean if the reset starts right AFTER we checked for the flag value then we still have a problem.

I recently worked (and still do) on device unplug support, I was thinking - can we use something alike drm_dev_enter/drm_dev_exit and drm_dev_unplug for our cause and this would replace both adev->in_gpu_reset and the lock_reset mutex(or rw_semaphote) ? In our case we also would need a  'drm_dev_plug_back' function too. Maybe we can even use the existing drm helpers as is as because while device is being reset it's effectively the same as if it's unplugged i think...

Some more bellow

> Signed-off-by: Dennis Li <Dennis.Li@xxxxxxx>
> Change-Id: I7f77a72795462587ed7d5f51fe53a594a0f1f708
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> index 80f32b3beb88..f235492799d7 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> @@ -963,9 +963,9 @@ struct amdgpu_device {
>   	bool                            in_suspend;
>   	bool				in_hibernate;
>   
> -	bool                            in_gpu_reset;
> +	atomic_t                        in_gpu_reset;
>   	enum pp_mp1_state               mp1_state;
> -	struct mutex  lock_reset;
> +	struct rw_semaphore	reset_sem;
>   	struct amdgpu_doorbell_index doorbell_index;
>   
>   	struct mutex			notifier_lock;
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c
> index 691c89705bcd..af71d8e93081 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c
> @@ -543,7 +543,7 @@ static int kgd_hqd_destroy(struct kgd_dev *kgd, void *mqd,
>   	uint32_t temp;
>   	struct v10_compute_mqd *m = get_mqd(mqd);
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EIO;
>   
>   #if 0
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c
> index 0b7e78748540..750a8308c868 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c
> @@ -425,7 +425,7 @@ static int kgd_hqd_destroy(struct kgd_dev *kgd, void *mqd,
>   	unsigned long flags, end_jiffies;
>   	int retry;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EIO;
>   
>   	acquire_queue(kgd, pipe_id, queue_id);
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c
> index ccd635b812b5..027793e0c1ca 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c
> @@ -421,7 +421,7 @@ static int kgd_hqd_destroy(struct kgd_dev *kgd, void *mqd,
>   	int retry;
>   	struct vi_mqd *m = get_mqd(mqd);
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EIO;
>   
>   	acquire_queue(kgd, pipe_id, queue_id);
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c
> index df841c2ac5e7..e4a77f7a4c2d 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c
> @@ -541,7 +541,7 @@ int kgd_gfx_v9_hqd_destroy(struct kgd_dev *kgd, void *mqd,
>   	uint32_t temp;
>   	struct v9_mqd *m = get_mqd(mqd);
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EIO;
>   
>   	acquire_queue(kgd, pipe_id, queue_id);
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> index ffbcaf4bfb8b..a94b3f862fc2 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> @@ -1292,6 +1292,8 @@ int amdgpu_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
>   	parser.adev = adev;
>   	parser.filp = filp;
>   
> +	down_read(&adev->reset_sem);
> +
>   	r = amdgpu_cs_parser_init(&parser, data);
>   	if (r) {
>   		DRM_ERROR("Failed to initialize parser %d!\n", r);
> @@ -1331,6 +1333,8 @@ int amdgpu_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
>   out:
>   	amdgpu_cs_parser_fini(&parser, r, reserved_buffers);
>   
> +	up_read(&adev->reset_sem);
> +
>   	return r;
>   }
>   
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
> index 8842c55d4490..d85d13f7a043 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
> @@ -358,6 +358,8 @@ static int amdgpu_ctx_query2(struct amdgpu_device *adev,
>   	if (atomic_read(&ctx->guilty))
>   		out->state.flags |= AMDGPU_CTX_QUERY2_FLAGS_GUILTY;
>   
> +	down_read(&adev->reset_sem);
> +
>   	/*query ue count*/
>   	ras_counter = amdgpu_ras_query_error_count(adev, false);
>   	/*ras counter is monotonic increasing*/
> @@ -373,6 +375,8 @@ static int amdgpu_ctx_query2(struct amdgpu_device *adev,
>   		ctx->ras_counter_ce = ras_counter;
>   	}
>   
> +	up_read(&adev->reset_sem);
> +
>   	mutex_unlock(&mgr->lock);
>   	return 0;
>   }
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
> index aeada7c9fbea..ecbff532a21b 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
> @@ -100,14 +100,14 @@ static int amdgpu_debugfs_autodump_open(struct inode *inode, struct file *file)
>   
>   	file->private_data = adev;
>   
> -	mutex_lock(&adev->lock_reset);
> +	down_read(&adev->reset_sem);
>   	if (adev->autodump.dumping.done) {
>   		reinit_completion(&adev->autodump.dumping);
>   		ret = 0;
>   	} else {
>   		ret = -EBUSY;
>   	}
> -	mutex_unlock(&adev->lock_reset);
> +	up_read(&adev->reset_sem);
>   
>   	return ret;
>   }
> @@ -126,7 +126,7 @@ static unsigned int amdgpu_debugfs_autodump_poll(struct file *file, struct poll_
>   
>   	poll_wait(file, &adev->autodump.gpu_hang, poll_table);
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return POLLIN | POLLRDNORM | POLLWRNORM;
>   
>   	return 0;
> @@ -1188,7 +1188,7 @@ static int amdgpu_debugfs_test_ib(struct seq_file *m, void *data)
>   	}
>   
>   	/* Avoid accidently unparking the sched thread during GPU reset */
> -	mutex_lock(&adev->lock_reset);
> +	down_read(&adev->reset_sem);
>   
>   	/* hold on the scheduler */
>   	for (i = 0; i < AMDGPU_MAX_RINGS; i++) {
> @@ -1215,7 +1215,7 @@ static int amdgpu_debugfs_test_ib(struct seq_file *m, void *data)
>   		kthread_unpark(ring->sched.thread);
>   	}
>   
> -	mutex_unlock(&adev->lock_reset);
> +	up_read(&adev->reset_sem);
>   
>   	pm_runtime_mark_last_busy(dev->dev);
>   	pm_runtime_put_autosuspend(dev->dev);
> @@ -1395,7 +1395,7 @@ static int amdgpu_debugfs_ib_preempt(void *data, u64 val)
>   		return -ENOMEM;
>   
>   	/* Avoid accidently unparking the sched thread during GPU reset */
> -	mutex_lock(&adev->lock_reset);
> +	down_read(&adev->reset_sem);
>   
>   	/* stop the scheduler */
>   	kthread_park(ring->sched.thread);
> @@ -1436,7 +1436,7 @@ static int amdgpu_debugfs_ib_preempt(void *data, u64 val)
>   	/* restart the scheduler */
>   	kthread_unpark(ring->sched.thread);
>   
> -	mutex_unlock(&adev->lock_reset);
> +	up_read(&adev->reset_sem);
>   
>   	ttm_bo_unlock_delayed_workqueue(&adev->mman.bdev, resched);
>   
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> index 2913e41f63a5..4f3ec9746d0c 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> @@ -1929,7 +1929,7 @@ static int amdgpu_device_fw_loading(struct amdgpu_device *adev)
>   			if (adev->ip_blocks[i].status.hw == true)
>   				break;
>   
> -			if (adev->in_gpu_reset || adev->in_suspend) {
> +			if (atomic_read(&adev->in_gpu_reset) || adev->in_suspend) {
>   				r = adev->ip_blocks[i].version->funcs->resume(adev);
>   				if (r) {
>   					DRM_ERROR("resume of IP block <%s> failed %d\n",
> @@ -2100,7 +2100,7 @@ static bool amdgpu_device_check_vram_lost(struct amdgpu_device *adev)
>   			AMDGPU_RESET_MAGIC_NUM))
>   		return true;
>   
> -	if (!adev->in_gpu_reset)
> +	if (!atomic_read(&adev->in_gpu_reset))
>   		return false;
>   
>   	/*
> @@ -3026,7 +3026,8 @@ int amdgpu_device_init(struct amdgpu_device *adev,
>   	mutex_init(&adev->mn_lock);
>   	mutex_init(&adev->virt.vf_errors.lock);
>   	hash_init(adev->mn_hash);
> -	mutex_init(&adev->lock_reset);
> +	init_rwsem(&adev->reset_sem);
> +	atomic_set(&adev->in_gpu_reset, 0);
>   	mutex_init(&adev->psp.mutex);
>   	mutex_init(&adev->notifier_lock);
>   
> @@ -4131,16 +4132,14 @@ static int amdgpu_do_asic_reset(struct amdgpu_hive_info *hive,
>   	return r;
>   }
>   
> -static bool amdgpu_device_lock_adev(struct amdgpu_device *adev, bool trylock)
> +static bool amdgpu_device_lock_adev(struct amdgpu_device *adev)
>   {
> -	if (trylock) {
> -		if (!mutex_trylock(&adev->lock_reset))
> -			return false;
> -	} else
> -		mutex_lock(&adev->lock_reset);
> +	if (atomic_cmpxchg(&adev->in_gpu_reset, 0, 1) != 0)
> +		return false;
> +
> +	down_write(&adev->reset_sem);


In V1 Christian raised a concern of this change increasing  the chance for gpu 
recovery abort due to collision of amdgpu_device_lock_adev doing down_write 
trylock with read side lock of any of the many operations here. I understand the 
need for this but before the 'trylock'  was used only for XGMI case because we 
knew XGMI hang on once card will trigger hangs on all other cards in the hive 
but now the 'trylock' logic (implemented by atomic_cmpxchg) also applies to the 
more common non XGMI case

Andrey


>   
>   	atomic_inc(&adev->gpu_reset_counter);
> -	adev->in_gpu_reset = true;
>   	switch (amdgpu_asic_reset_method(adev)) {
>   	case AMD_RESET_METHOD_MODE1:
>   		adev->mp1_state = PP_MP1_STATE_SHUTDOWN;
> @@ -4160,8 +4159,8 @@ static void amdgpu_device_unlock_adev(struct amdgpu_device *adev)
>   {
>   	amdgpu_vf_error_trans_all(adev);
>   	adev->mp1_state = PP_MP1_STATE_NONE;
> -	adev->in_gpu_reset = false;
> -	mutex_unlock(&adev->lock_reset);
> +	atomic_set(&adev->in_gpu_reset, 0);
> +	up_write(&adev->reset_sem);
>   }
>   
>   static void amdgpu_device_resume_display_audio(struct amdgpu_device *adev)
> @@ -4297,7 +4296,7 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
>   
>   	/* block all schedulers and reset given job's ring */
>   	list_for_each_entry(tmp_adev, device_list_handle, gmc.xgmi.head) {
> -		if (!amdgpu_device_lock_adev(tmp_adev, !hive)) {
> +		if (!amdgpu_device_lock_adev(tmp_adev)) {
>   			DRM_INFO("Bailing on TDR for s_job:%llx, as another already in progress",
>   				  job ? job->base.id : -1);
>   			mutex_unlock(&hive->hive_lock);
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
> index 77d988a0033f..96df3669a9d2 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
> @@ -670,6 +670,8 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
>   		bo_va = NULL;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	switch (args->operation) {
>   	case AMDGPU_VA_OP_MAP:
>   		va_flags = amdgpu_gem_va_map_flags(adev, args->flags);
> @@ -699,6 +701,8 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
>   		amdgpu_gem_va_update_vm(adev, &fpriv->vm, bo_va,
>   					args->operation);
>   
> +	up_read(&adev->reset_sem);
> +
>   error_backoff:
>   	ttm_eu_backoff_reservation(&ticket, &list);
>   
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
> index d612033a23ac..c9f13f836a7a 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
> @@ -710,7 +710,7 @@ uint32_t amdgpu_kiq_rreg(struct amdgpu_device *adev, uint32_t reg)
>   	 *
>   	 * also don't wait anymore for IRQ context
>   	 * */
> -	if (r < 1 && (adev->in_gpu_reset || in_interrupt()))
> +	if (r < 1 && (atomic_read(&adev->in_gpu_reset) || in_interrupt()))
>   		goto failed_kiq_read;
>   
>   	might_sleep();
> @@ -768,7 +768,7 @@ void amdgpu_kiq_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v)
>   	 *
>   	 * also don't wait anymore for IRQ context
>   	 * */
> -	if (r < 1 && (adev->in_gpu_reset || in_interrupt()))
> +	if (r < 1 && (atomic_read(&adev->in_gpu_reset) || in_interrupt()))
>   		goto failed_kiq_write;
>   
>   	might_sleep();
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
> index 937029ad5271..367528de1658 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
> @@ -226,8 +226,10 @@ static struct dma_fence *amdgpu_job_run(struct drm_sched_job *sched_job)
>   	if (finished->error < 0) {
>   		DRM_INFO("Skip scheduling IBs!\n");
>   	} else {
> +		down_read(&ring->adev->reset_sem);
>   		r = amdgpu_ib_schedule(ring, job->num_ibs, job->ibs, job,
>   				       &fence);
> +		up_read(&ring->adev->reset_sem);
>   		if (r)
>   			DRM_ERROR("Error scheduling IBs (%d)\n", r);
>   	}
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
> index 9c2c5e4ed1eb..eebbe2103e32 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
> @@ -1084,6 +1084,8 @@ void amdgpu_driver_postclose_kms(struct drm_device *dev,
>   	if (!fpriv)
>   		return;
>   
> +	down_read(&adev->reset_sem);
> +
>   	pm_runtime_get_sync(dev->dev);
>   
>   	if (amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_UVD) != NULL)
> @@ -1122,6 +1124,8 @@ void amdgpu_driver_postclose_kms(struct drm_device *dev,
>   
>   	pm_runtime_mark_last_busy(dev->dev);
>   	pm_runtime_put_autosuspend(dev->dev);
> +
> +	up_read(&adev->reset_sem);
>   }
>   
>   /*
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c
> index d2401379bd33..61b23ae906d7 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c
> @@ -163,7 +163,7 @@ static ssize_t amdgpu_get_power_dpm_state(struct device *dev,
>   	enum amd_pm_state_type pm;
>   	int ret;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = pm_runtime_get_sync(ddev->dev);
> @@ -172,6 +172,8 @@ static ssize_t amdgpu_get_power_dpm_state(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev)) {
>   		if (adev->smu.ppt_funcs->get_current_power_state)
>   			pm = smu_get_current_power_state(&adev->smu);
> @@ -183,6 +185,8 @@ static ssize_t amdgpu_get_power_dpm_state(struct device *dev,
>   		pm = adev->pm.dpm.user_state;
>   	}
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> @@ -201,7 +205,7 @@ static ssize_t amdgpu_set_power_dpm_state(struct device *dev,
>   	enum amd_pm_state_type  state;
>   	int ret;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	if (strncmp("battery", buf, strlen("battery")) == 0)
> @@ -219,6 +223,8 @@ static ssize_t amdgpu_set_power_dpm_state(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev)) {
>   		mutex_lock(&adev->pm.mutex);
>   		adev->pm.dpm.user_state = state;
> @@ -232,6 +238,9 @@ static ssize_t amdgpu_set_power_dpm_state(struct device *dev,
>   
>   		amdgpu_pm_compute_clocks(adev);
>   	}
> +
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> @@ -307,7 +316,7 @@ static ssize_t amdgpu_get_power_dpm_force_performance_level(struct device *dev,
>   	enum amd_dpm_forced_level level = 0xff;
>   	int ret;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = pm_runtime_get_sync(ddev->dev);
> @@ -316,6 +325,8 @@ static ssize_t amdgpu_get_power_dpm_force_performance_level(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev))
>   		level = smu_get_performance_level(&adev->smu);
>   	else if (adev->powerplay.pp_funcs->get_performance_level)
> @@ -323,6 +334,8 @@ static ssize_t amdgpu_get_power_dpm_force_performance_level(struct device *dev,
>   	else
>   		level = adev->pm.dpm.forced_level;
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> @@ -349,7 +362,7 @@ static ssize_t amdgpu_set_power_dpm_force_performance_level(struct device *dev,
>   	enum amd_dpm_forced_level current_level = 0xff;
>   	int ret = 0;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	if (strncmp("low", buf, strlen("low")) == 0) {
> @@ -380,6 +393,8 @@ static ssize_t amdgpu_set_power_dpm_force_performance_level(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev))
>   		current_level = smu_get_performance_level(&adev->smu);
>   	else if (adev->powerplay.pp_funcs->get_performance_level)
> @@ -388,7 +403,8 @@ static ssize_t amdgpu_set_power_dpm_force_performance_level(struct device *dev,
>   	if (current_level == level) {
>   		pm_runtime_mark_last_busy(ddev->dev);
>   		pm_runtime_put_autosuspend(ddev->dev);
> -		return count;
> +		ret = count;
> +		goto pro_end;
>   	}
>   
>   	if (adev->asic_type == CHIP_RAVEN) {
> @@ -409,7 +425,8 @@ static ssize_t amdgpu_set_power_dpm_force_performance_level(struct device *dev,
>   		pr_err("Currently not in any profile mode!\n");
>   		pm_runtime_mark_last_busy(ddev->dev);
>   		pm_runtime_put_autosuspend(ddev->dev);
> -		return -EINVAL;
> +		ret = -EINVAL;
> +		goto pro_end;
>   	}
>   
>   	if (is_support_sw_smu(adev)) {
> @@ -417,7 +434,8 @@ static ssize_t amdgpu_set_power_dpm_force_performance_level(struct device *dev,
>   		if (ret) {
>   			pm_runtime_mark_last_busy(ddev->dev);
>   			pm_runtime_put_autosuspend(ddev->dev);
> -			return -EINVAL;
> +			ret = -EINVAL;
> +			goto pro_end;
>   		}
>   	} else if (adev->powerplay.pp_funcs->force_performance_level) {
>   		mutex_lock(&adev->pm.mutex);
> @@ -425,14 +443,16 @@ static ssize_t amdgpu_set_power_dpm_force_performance_level(struct device *dev,
>   			mutex_unlock(&adev->pm.mutex);
>   			pm_runtime_mark_last_busy(ddev->dev);
>   			pm_runtime_put_autosuspend(ddev->dev);
> -			return -EINVAL;
> +			ret = -EINVAL;
> +			goto pro_end;
>   		}
>   		ret = amdgpu_dpm_force_performance_level(adev, level);
>   		if (ret) {
>   			mutex_unlock(&adev->pm.mutex);
>   			pm_runtime_mark_last_busy(ddev->dev);
>   			pm_runtime_put_autosuspend(ddev->dev);
> -			return -EINVAL;
> +			ret = -EINVAL;
> +			goto pro_end;
>   		} else {
>   			adev->pm.dpm.forced_level = level;
>   		}
> @@ -441,7 +461,9 @@ static ssize_t amdgpu_set_power_dpm_force_performance_level(struct device *dev,
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> -	return count;
> +pro_end:
> +	up_read(&adev->reset_sem);
> +	return ret;
>   }
>   
>   static ssize_t amdgpu_get_pp_num_states(struct device *dev,
> @@ -453,7 +475,7 @@ static ssize_t amdgpu_get_pp_num_states(struct device *dev,
>   	struct pp_states_info data;
>   	int i, buf_len, ret;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = pm_runtime_get_sync(ddev->dev);
> @@ -497,7 +519,7 @@ static ssize_t amdgpu_get_pp_cur_state(struct device *dev,
>   	enum amd_pm_state_type pm = 0;
>   	int i = 0, ret = 0;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = pm_runtime_get_sync(ddev->dev);
> @@ -538,7 +560,7 @@ static ssize_t amdgpu_get_pp_force_state(struct device *dev,
>   	struct drm_device *ddev = dev_get_drvdata(dev);
>   	struct amdgpu_device *adev = ddev->dev_private;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	if (adev->pp_force_state_enabled)
> @@ -558,7 +580,7 @@ static ssize_t amdgpu_set_pp_force_state(struct device *dev,
>   	unsigned long idx;
>   	int ret;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	if (strlen(buf) == 1)
> @@ -584,6 +606,7 @@ static ssize_t amdgpu_set_pp_force_state(struct device *dev,
>   			return ret;
>   		}
>   
> +		down_read(&adev->reset_sem);
>   		/* only set user selected power states */
>   		if (state != POWER_STATE_TYPE_INTERNAL_BOOT &&
>   		    state != POWER_STATE_TYPE_DEFAULT) {
> @@ -591,6 +614,8 @@ static ssize_t amdgpu_set_pp_force_state(struct device *dev,
>   					AMD_PP_TASK_ENABLE_USER_STATE, &state);
>   			adev->pp_force_state_enabled = true;
>   		}
> +		up_read(&adev->reset_sem);
> +
>   		pm_runtime_mark_last_busy(ddev->dev);
>   		pm_runtime_put_autosuspend(ddev->dev);
>   	}
> @@ -618,7 +643,7 @@ static ssize_t amdgpu_get_pp_table(struct device *dev,
>   	char *table = NULL;
>   	int size, ret;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = pm_runtime_get_sync(ddev->dev);
> @@ -662,7 +687,7 @@ static ssize_t amdgpu_set_pp_table(struct device *dev,
>   	struct amdgpu_device *adev = ddev->dev_private;
>   	int ret = 0;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = pm_runtime_get_sync(ddev->dev);
> @@ -671,16 +696,21 @@ static ssize_t amdgpu_set_pp_table(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev)) {
>   		ret = smu_sys_set_pp_table(&adev->smu, (void *)buf, count);
>   		if (ret) {
>   			pm_runtime_mark_last_busy(ddev->dev);
>   			pm_runtime_put_autosuspend(ddev->dev);
> +			up_read(&adev->reset_sem);
>   			return ret;
>   		}
>   	} else if (adev->powerplay.pp_funcs->set_pp_table)
>   		amdgpu_dpm_set_pp_table(adev, buf, count);
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> @@ -769,7 +799,7 @@ static ssize_t amdgpu_set_pp_od_clk_voltage(struct device *dev,
>   	const char delimiter[3] = {' ', '\n', '\0'};
>   	uint32_t type;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	if (count > 127)
> @@ -813,6 +843,10 @@ static ssize_t amdgpu_set_pp_od_clk_voltage(struct device *dev,
>   		return ret;
>   	}
>   
> +	ret = count;
> +
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev)) {
>   		ret = smu_od_edit_dpm_table(&adev->smu, type,
>   					    parameter, parameter_size);
> @@ -820,7 +854,8 @@ static ssize_t amdgpu_set_pp_od_clk_voltage(struct device *dev,
>   		if (ret) {
>   			pm_runtime_mark_last_busy(ddev->dev);
>   			pm_runtime_put_autosuspend(ddev->dev);
> -			return -EINVAL;
> +			ret = -EINVAL;
> +			goto pro_end;
>   		}
>   	} else {
>   		if (adev->powerplay.pp_funcs->odn_edit_dpm_table) {
> @@ -829,7 +864,8 @@ static ssize_t amdgpu_set_pp_od_clk_voltage(struct device *dev,
>   			if (ret) {
>   				pm_runtime_mark_last_busy(ddev->dev);
>   				pm_runtime_put_autosuspend(ddev->dev);
> -				return -EINVAL;
> +				ret = -EINVAL;
> +				goto pro_end;
>   			}
>   		}
>   
> @@ -840,18 +876,22 @@ static ssize_t amdgpu_set_pp_od_clk_voltage(struct device *dev,
>   						NULL);
>   				pm_runtime_mark_last_busy(ddev->dev);
>   				pm_runtime_put_autosuspend(ddev->dev);
> -				return count;
> +				ret = count;
> +				goto pro_end;
>   			} else {
>   				pm_runtime_mark_last_busy(ddev->dev);
>   				pm_runtime_put_autosuspend(ddev->dev);
> -				return -EINVAL;
> +				ret = -EINVAL;
> +				goto pro_end;
>   			}
>   		}
>   	}
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> -	return count;
> +pro_end:
> +	up_read(&adev->reset_sem);
> +	return ret;
>   }
>   
>   static ssize_t amdgpu_get_pp_od_clk_voltage(struct device *dev,
> @@ -863,7 +903,7 @@ static ssize_t amdgpu_get_pp_od_clk_voltage(struct device *dev,
>   	ssize_t size;
>   	int ret;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = pm_runtime_get_sync(ddev->dev);
> @@ -917,7 +957,7 @@ static ssize_t amdgpu_set_pp_features(struct device *dev,
>   	uint64_t featuremask;
>   	int ret;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = kstrtou64(buf, 0, &featuremask);
> @@ -932,11 +972,13 @@ static ssize_t amdgpu_set_pp_features(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
>   	if (is_support_sw_smu(adev)) {
>   		ret = smu_sys_set_pp_feature_mask(&adev->smu, featuremask);
>   		if (ret) {
>   			pm_runtime_mark_last_busy(ddev->dev);
>   			pm_runtime_put_autosuspend(ddev->dev);
> +			up_read(&adev->reset_sem);
>   			return -EINVAL;
>   		}
>   	} else if (adev->powerplay.pp_funcs->set_ppfeature_status) {
> @@ -944,9 +986,12 @@ static ssize_t amdgpu_set_pp_features(struct device *dev,
>   		if (ret) {
>   			pm_runtime_mark_last_busy(ddev->dev);
>   			pm_runtime_put_autosuspend(ddev->dev);
> +			up_read(&adev->reset_sem);
>   			return -EINVAL;
>   		}
>   	}
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> @@ -962,7 +1007,7 @@ static ssize_t amdgpu_get_pp_features(struct device *dev,
>   	ssize_t size;
>   	int ret;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = pm_runtime_get_sync(ddev->dev);
> @@ -971,6 +1016,8 @@ static ssize_t amdgpu_get_pp_features(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev))
>   		size = smu_sys_get_pp_feature_mask(&adev->smu, buf);
>   	else if (adev->powerplay.pp_funcs->get_ppfeature_status)
> @@ -978,6 +1025,8 @@ static ssize_t amdgpu_get_pp_features(struct device *dev,
>   	else
>   		size = snprintf(buf, PAGE_SIZE, "\n");
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> @@ -1023,7 +1072,7 @@ static ssize_t amdgpu_get_pp_dpm_sclk(struct device *dev,
>   	ssize_t size;
>   	int ret;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = pm_runtime_get_sync(ddev->dev);
> @@ -1032,6 +1081,8 @@ static ssize_t amdgpu_get_pp_dpm_sclk(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev))
>   		size = smu_print_clk_levels(&adev->smu, SMU_SCLK, buf);
>   	else if (adev->powerplay.pp_funcs->print_clock_levels)
> @@ -1039,6 +1090,8 @@ static ssize_t amdgpu_get_pp_dpm_sclk(struct device *dev,
>   	else
>   		size = snprintf(buf, PAGE_SIZE, "\n");
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> @@ -1091,7 +1144,7 @@ static ssize_t amdgpu_set_pp_dpm_sclk(struct device *dev,
>   	int ret;
>   	uint32_t mask = 0;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = amdgpu_read_mask(buf, count, &mask);
> @@ -1104,11 +1157,15 @@ static ssize_t amdgpu_set_pp_dpm_sclk(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev))
>   		ret = smu_force_clk_levels(&adev->smu, SMU_SCLK, mask, true);
>   	else if (adev->powerplay.pp_funcs->force_clock_level)
>   		ret = amdgpu_dpm_force_clock_level(adev, PP_SCLK, mask);
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> @@ -1127,7 +1184,7 @@ static ssize_t amdgpu_get_pp_dpm_mclk(struct device *dev,
>   	ssize_t size;
>   	int ret;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = pm_runtime_get_sync(ddev->dev);
> @@ -1136,6 +1193,8 @@ static ssize_t amdgpu_get_pp_dpm_mclk(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev))
>   		size = smu_print_clk_levels(&adev->smu, SMU_MCLK, buf);
>   	else if (adev->powerplay.pp_funcs->print_clock_levels)
> @@ -1143,6 +1202,8 @@ static ssize_t amdgpu_get_pp_dpm_mclk(struct device *dev,
>   	else
>   		size = snprintf(buf, PAGE_SIZE, "\n");
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> @@ -1159,7 +1220,7 @@ static ssize_t amdgpu_set_pp_dpm_mclk(struct device *dev,
>   	uint32_t mask = 0;
>   	int ret;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = amdgpu_read_mask(buf, count, &mask);
> @@ -1172,11 +1233,15 @@ static ssize_t amdgpu_set_pp_dpm_mclk(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev))
>   		ret = smu_force_clk_levels(&adev->smu, SMU_MCLK, mask, true);
>   	else if (adev->powerplay.pp_funcs->force_clock_level)
>   		ret = amdgpu_dpm_force_clock_level(adev, PP_MCLK, mask);
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> @@ -1195,7 +1260,7 @@ static ssize_t amdgpu_get_pp_dpm_socclk(struct device *dev,
>   	ssize_t size;
>   	int ret;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = pm_runtime_get_sync(ddev->dev);
> @@ -1204,6 +1269,8 @@ static ssize_t amdgpu_get_pp_dpm_socclk(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev))
>   		size = smu_print_clk_levels(&adev->smu, SMU_SOCCLK, buf);
>   	else if (adev->powerplay.pp_funcs->print_clock_levels)
> @@ -1211,6 +1278,8 @@ static ssize_t amdgpu_get_pp_dpm_socclk(struct device *dev,
>   	else
>   		size = snprintf(buf, PAGE_SIZE, "\n");
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> @@ -1227,7 +1296,7 @@ static ssize_t amdgpu_set_pp_dpm_socclk(struct device *dev,
>   	int ret;
>   	uint32_t mask = 0;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = amdgpu_read_mask(buf, count, &mask);
> @@ -1240,6 +1309,8 @@ static ssize_t amdgpu_set_pp_dpm_socclk(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev))
>   		ret = smu_force_clk_levels(&adev->smu, SMU_SOCCLK, mask, true);
>   	else if (adev->powerplay.pp_funcs->force_clock_level)
> @@ -1247,6 +1318,8 @@ static ssize_t amdgpu_set_pp_dpm_socclk(struct device *dev,
>   	else
>   		ret = 0;
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> @@ -1265,7 +1338,7 @@ static ssize_t amdgpu_get_pp_dpm_fclk(struct device *dev,
>   	ssize_t size;
>   	int ret;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = pm_runtime_get_sync(ddev->dev);
> @@ -1274,6 +1347,8 @@ static ssize_t amdgpu_get_pp_dpm_fclk(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev))
>   		size = smu_print_clk_levels(&adev->smu, SMU_FCLK, buf);
>   	else if (adev->powerplay.pp_funcs->print_clock_levels)
> @@ -1281,6 +1356,8 @@ static ssize_t amdgpu_get_pp_dpm_fclk(struct device *dev,
>   	else
>   		size = snprintf(buf, PAGE_SIZE, "\n");
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> @@ -1297,7 +1374,7 @@ static ssize_t amdgpu_set_pp_dpm_fclk(struct device *dev,
>   	int ret;
>   	uint32_t mask = 0;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = amdgpu_read_mask(buf, count, &mask);
> @@ -1310,6 +1387,8 @@ static ssize_t amdgpu_set_pp_dpm_fclk(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev))
>   		ret = smu_force_clk_levels(&adev->smu, SMU_FCLK, mask, true);
>   	else if (adev->powerplay.pp_funcs->force_clock_level)
> @@ -1317,6 +1396,8 @@ static ssize_t amdgpu_set_pp_dpm_fclk(struct device *dev,
>   	else
>   		ret = 0;
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> @@ -1335,7 +1416,7 @@ static ssize_t amdgpu_get_pp_dpm_dcefclk(struct device *dev,
>   	ssize_t size;
>   	int ret;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = pm_runtime_get_sync(ddev->dev);
> @@ -1344,6 +1425,8 @@ static ssize_t amdgpu_get_pp_dpm_dcefclk(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev))
>   		size = smu_print_clk_levels(&adev->smu, SMU_DCEFCLK, buf);
>   	else if (adev->powerplay.pp_funcs->print_clock_levels)
> @@ -1351,6 +1434,8 @@ static ssize_t amdgpu_get_pp_dpm_dcefclk(struct device *dev,
>   	else
>   		size = snprintf(buf, PAGE_SIZE, "\n");
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> @@ -1367,7 +1452,7 @@ static ssize_t amdgpu_set_pp_dpm_dcefclk(struct device *dev,
>   	int ret;
>   	uint32_t mask = 0;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = amdgpu_read_mask(buf, count, &mask);
> @@ -1380,6 +1465,8 @@ static ssize_t amdgpu_set_pp_dpm_dcefclk(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev))
>   		ret = smu_force_clk_levels(&adev->smu, SMU_DCEFCLK, mask, true);
>   	else if (adev->powerplay.pp_funcs->force_clock_level)
> @@ -1387,6 +1474,8 @@ static ssize_t amdgpu_set_pp_dpm_dcefclk(struct device *dev,
>   	else
>   		ret = 0;
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> @@ -1405,7 +1494,7 @@ static ssize_t amdgpu_get_pp_dpm_pcie(struct device *dev,
>   	ssize_t size;
>   	int ret;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = pm_runtime_get_sync(ddev->dev);
> @@ -1414,6 +1503,8 @@ static ssize_t amdgpu_get_pp_dpm_pcie(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev))
>   		size = smu_print_clk_levels(&adev->smu, SMU_PCIE, buf);
>   	else if (adev->powerplay.pp_funcs->print_clock_levels)
> @@ -1421,6 +1512,8 @@ static ssize_t amdgpu_get_pp_dpm_pcie(struct device *dev,
>   	else
>   		size = snprintf(buf, PAGE_SIZE, "\n");
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> @@ -1437,7 +1530,7 @@ static ssize_t amdgpu_set_pp_dpm_pcie(struct device *dev,
>   	int ret;
>   	uint32_t mask = 0;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = amdgpu_read_mask(buf, count, &mask);
> @@ -1450,6 +1543,8 @@ static ssize_t amdgpu_set_pp_dpm_pcie(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev))
>   		ret = smu_force_clk_levels(&adev->smu, SMU_PCIE, mask, true);
>   	else if (adev->powerplay.pp_funcs->force_clock_level)
> @@ -1457,6 +1552,8 @@ static ssize_t amdgpu_set_pp_dpm_pcie(struct device *dev,
>   	else
>   		ret = 0;
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> @@ -1475,7 +1572,7 @@ static ssize_t amdgpu_get_pp_sclk_od(struct device *dev,
>   	uint32_t value = 0;
>   	int ret;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = pm_runtime_get_sync(ddev->dev);
> @@ -1484,11 +1581,15 @@ static ssize_t amdgpu_get_pp_sclk_od(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev))
>   		value = smu_get_od_percentage(&(adev->smu), SMU_OD_SCLK);
>   	else if (adev->powerplay.pp_funcs->get_sclk_od)
>   		value = amdgpu_dpm_get_sclk_od(adev);
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> @@ -1505,7 +1606,7 @@ static ssize_t amdgpu_set_pp_sclk_od(struct device *dev,
>   	int ret;
>   	long int value;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = kstrtol(buf, 0, &value);
> @@ -1519,6 +1620,8 @@ static ssize_t amdgpu_set_pp_sclk_od(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev)) {
>   		value = smu_set_od_percentage(&(adev->smu), SMU_OD_SCLK, (uint32_t)value);
>   	} else {
> @@ -1533,6 +1636,8 @@ static ssize_t amdgpu_set_pp_sclk_od(struct device *dev,
>   		}
>   	}
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> @@ -1548,7 +1653,7 @@ static ssize_t amdgpu_get_pp_mclk_od(struct device *dev,
>   	uint32_t value = 0;
>   	int ret;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = pm_runtime_get_sync(ddev->dev);
> @@ -1557,11 +1662,15 @@ static ssize_t amdgpu_get_pp_mclk_od(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev))
>   		value = smu_get_od_percentage(&(adev->smu), SMU_OD_MCLK);
>   	else if (adev->powerplay.pp_funcs->get_mclk_od)
>   		value = amdgpu_dpm_get_mclk_od(adev);
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> @@ -1578,7 +1687,7 @@ static ssize_t amdgpu_set_pp_mclk_od(struct device *dev,
>   	int ret;
>   	long int value;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = kstrtol(buf, 0, &value);
> @@ -1592,6 +1701,8 @@ static ssize_t amdgpu_set_pp_mclk_od(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev)) {
>   		value = smu_set_od_percentage(&(adev->smu), SMU_OD_MCLK, (uint32_t)value);
>   	} else {
> @@ -1606,6 +1717,8 @@ static ssize_t amdgpu_set_pp_mclk_od(struct device *dev,
>   		}
>   	}
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> @@ -1641,7 +1754,7 @@ static ssize_t amdgpu_get_pp_power_profile_mode(struct device *dev,
>   	ssize_t size;
>   	int ret;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = pm_runtime_get_sync(ddev->dev);
> @@ -1650,6 +1763,8 @@ static ssize_t amdgpu_get_pp_power_profile_mode(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev))
>   		size = smu_get_power_profile_mode(&adev->smu, buf);
>   	else if (adev->powerplay.pp_funcs->get_power_profile_mode)
> @@ -1657,6 +1772,8 @@ static ssize_t amdgpu_get_pp_power_profile_mode(struct device *dev,
>   	else
>   		size = snprintf(buf, PAGE_SIZE, "\n");
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> @@ -1681,7 +1798,7 @@ static ssize_t amdgpu_set_pp_power_profile_mode(struct device *dev,
>   	long int profile_mode = 0;
>   	const char delimiter[3] = {' ', '\n', '\0'};
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	tmp[0] = *(buf);
> @@ -1715,11 +1832,15 @@ static ssize_t amdgpu_set_pp_power_profile_mode(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev))
>   		ret = smu_set_power_profile_mode(&adev->smu, parameter, parameter_size, true);
>   	else if (adev->powerplay.pp_funcs->set_power_profile_mode)
>   		ret = amdgpu_dpm_set_power_profile_mode(adev, parameter, parameter_size);
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> @@ -1745,7 +1866,7 @@ static ssize_t amdgpu_get_gpu_busy_percent(struct device *dev,
>   	struct amdgpu_device *adev = ddev->dev_private;
>   	int r, value, size = sizeof(value);
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	r = pm_runtime_get_sync(ddev->dev);
> @@ -1754,9 +1875,11 @@ static ssize_t amdgpu_get_gpu_busy_percent(struct device *dev,
>   		return r;
>   	}
>   
> +	down_read(&adev->reset_sem);
>   	/* read the IP busy sensor */
>   	r = amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_GPU_LOAD,
>   				   (void *)&value, &size);
> +	up_read(&adev->reset_sem);
>   
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
> @@ -1783,7 +1906,7 @@ static ssize_t amdgpu_get_mem_busy_percent(struct device *dev,
>   	struct amdgpu_device *adev = ddev->dev_private;
>   	int r, value, size = sizeof(value);
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	r = pm_runtime_get_sync(ddev->dev);
> @@ -1792,10 +1915,14 @@ static ssize_t amdgpu_get_mem_busy_percent(struct device *dev,
>   		return r;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	/* read the IP busy sensor */
>   	r = amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_MEM_LOAD,
>   				   (void *)&value, &size);
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> @@ -1826,7 +1953,7 @@ static ssize_t amdgpu_get_pcie_bw(struct device *dev,
>   	uint64_t count0 = 0, count1 = 0;
>   	int ret;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	if (adev->flags & AMD_IS_APU)
> @@ -1841,8 +1968,12 @@ static ssize_t amdgpu_get_pcie_bw(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	amdgpu_asic_get_pcie_usage(adev, &count0, &count1);
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(ddev->dev);
>   	pm_runtime_put_autosuspend(ddev->dev);
>   
> @@ -1867,7 +1998,7 @@ static ssize_t amdgpu_get_unique_id(struct device *dev,
>   	struct drm_device *ddev = dev_get_drvdata(dev);
>   	struct amdgpu_device *adev = ddev->dev_private;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	if (adev->unique_id)
> @@ -2134,7 +2265,7 @@ static ssize_t amdgpu_hwmon_show_temp(struct device *dev,
>   	int channel = to_sensor_dev_attr(attr)->index;
>   	int r, temp = 0, size = sizeof(temp);
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	if (channel >= PP_TEMP_MAX)
> @@ -2146,6 +2277,8 @@ static ssize_t amdgpu_hwmon_show_temp(struct device *dev,
>   		return r;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	switch (channel) {
>   	case PP_TEMP_JUNCTION:
>   		/* get current junction temperature */
> @@ -2167,6 +2300,8 @@ static ssize_t amdgpu_hwmon_show_temp(struct device *dev,
>   		break;
>   	}
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(adev->ddev->dev);
>   	pm_runtime_put_autosuspend(adev->ddev->dev);
>   
> @@ -2270,7 +2405,7 @@ static ssize_t amdgpu_hwmon_get_pwm1_enable(struct device *dev,
>   	u32 pwm_mode = 0;
>   	int ret;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = pm_runtime_get_sync(adev->ddev->dev);
> @@ -2279,18 +2414,23 @@ static ssize_t amdgpu_hwmon_get_pwm1_enable(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev)) {
>   		pwm_mode = smu_get_fan_control_mode(&adev->smu);
>   	} else {
>   		if (!adev->powerplay.pp_funcs->get_fan_control_mode) {
>   			pm_runtime_mark_last_busy(adev->ddev->dev);
>   			pm_runtime_put_autosuspend(adev->ddev->dev);
> +			up_read(&adev->reset_sem);
>   			return -EINVAL;
>   		}
>   
>   		pwm_mode = amdgpu_dpm_get_fan_control_mode(adev);
>   	}
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(adev->ddev->dev);
>   	pm_runtime_put_autosuspend(adev->ddev->dev);
>   
> @@ -2306,7 +2446,7 @@ static ssize_t amdgpu_hwmon_set_pwm1_enable(struct device *dev,
>   	int err, ret;
>   	int value;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	err = kstrtoint(buf, 10, &value);
> @@ -2319,18 +2459,23 @@ static ssize_t amdgpu_hwmon_set_pwm1_enable(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev)) {
>   		smu_set_fan_control_mode(&adev->smu, value);
>   	} else {
>   		if (!adev->powerplay.pp_funcs->set_fan_control_mode) {
>   			pm_runtime_mark_last_busy(adev->ddev->dev);
>   			pm_runtime_put_autosuspend(adev->ddev->dev);
> +			up_read(&adev->reset_sem);
>   			return -EINVAL;
>   		}
>   
>   		amdgpu_dpm_set_fan_control_mode(adev, value);
>   	}
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(adev->ddev->dev);
>   	pm_runtime_put_autosuspend(adev->ddev->dev);
>   
> @@ -2360,7 +2505,7 @@ static ssize_t amdgpu_hwmon_set_pwm1(struct device *dev,
>   	u32 value;
>   	u32 pwm_mode;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	err = pm_runtime_get_sync(adev->ddev->dev);
> @@ -2369,11 +2514,15 @@ static ssize_t amdgpu_hwmon_set_pwm1(struct device *dev,
>   		return err;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev))
>   		pwm_mode = smu_get_fan_control_mode(&adev->smu);
>   	else
>   		pwm_mode = amdgpu_dpm_get_fan_control_mode(adev);
>   
> +	up_read(&adev->reset_sem);
> +
>   	if (pwm_mode != AMD_FAN_CTRL_MANUAL) {
>   		pr_info("manual fan speed control should be enabled first\n");
>   		pm_runtime_mark_last_busy(adev->ddev->dev);
> @@ -2414,7 +2563,7 @@ static ssize_t amdgpu_hwmon_get_pwm1(struct device *dev,
>   	int err;
>   	u32 speed = 0;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	err = pm_runtime_get_sync(adev->ddev->dev);
> @@ -2423,6 +2572,8 @@ static ssize_t amdgpu_hwmon_get_pwm1(struct device *dev,
>   		return err;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev))
>   		err = smu_get_fan_speed_percent(&adev->smu, &speed);
>   	else if (adev->powerplay.pp_funcs->get_fan_speed_percent)
> @@ -2430,6 +2581,8 @@ static ssize_t amdgpu_hwmon_get_pwm1(struct device *dev,
>   	else
>   		err = -EINVAL;
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(adev->ddev->dev);
>   	pm_runtime_put_autosuspend(adev->ddev->dev);
>   
> @@ -2449,7 +2602,7 @@ static ssize_t amdgpu_hwmon_get_fan1_input(struct device *dev,
>   	int err;
>   	u32 speed = 0;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	err = pm_runtime_get_sync(adev->ddev->dev);
> @@ -2458,6 +2611,8 @@ static ssize_t amdgpu_hwmon_get_fan1_input(struct device *dev,
>   		return err;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev))
>   		err = smu_get_fan_speed_rpm(&adev->smu, &speed);
>   	else if (adev->powerplay.pp_funcs->get_fan_speed_rpm)
> @@ -2465,6 +2620,8 @@ static ssize_t amdgpu_hwmon_get_fan1_input(struct device *dev,
>   	else
>   		err = -EINVAL;
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(adev->ddev->dev);
>   	pm_runtime_put_autosuspend(adev->ddev->dev);
>   
> @@ -2483,7 +2640,7 @@ static ssize_t amdgpu_hwmon_get_fan1_min(struct device *dev,
>   	u32 size = sizeof(min_rpm);
>   	int r;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	r = pm_runtime_get_sync(adev->ddev->dev);
> @@ -2492,9 +2649,13 @@ static ssize_t amdgpu_hwmon_get_fan1_min(struct device *dev,
>   		return r;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	r = amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_MIN_FAN_RPM,
>   				   (void *)&min_rpm, &size);
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(adev->ddev->dev);
>   	pm_runtime_put_autosuspend(adev->ddev->dev);
>   
> @@ -2513,7 +2674,7 @@ static ssize_t amdgpu_hwmon_get_fan1_max(struct device *dev,
>   	u32 size = sizeof(max_rpm);
>   	int r;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	r = pm_runtime_get_sync(adev->ddev->dev);
> @@ -2522,9 +2683,13 @@ static ssize_t amdgpu_hwmon_get_fan1_max(struct device *dev,
>   		return r;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	r = amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_MAX_FAN_RPM,
>   				   (void *)&max_rpm, &size);
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(adev->ddev->dev);
>   	pm_runtime_put_autosuspend(adev->ddev->dev);
>   
> @@ -2542,7 +2707,7 @@ static ssize_t amdgpu_hwmon_get_fan1_target(struct device *dev,
>   	int err;
>   	u32 rpm = 0;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	err = pm_runtime_get_sync(adev->ddev->dev);
> @@ -2551,6 +2716,8 @@ static ssize_t amdgpu_hwmon_get_fan1_target(struct device *dev,
>   		return err;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev))
>   		err = smu_get_fan_speed_rpm(&adev->smu, &rpm);
>   	else if (adev->powerplay.pp_funcs->get_fan_speed_rpm)
> @@ -2558,6 +2725,8 @@ static ssize_t amdgpu_hwmon_get_fan1_target(struct device *dev,
>   	else
>   		err = -EINVAL;
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(adev->ddev->dev);
>   	pm_runtime_put_autosuspend(adev->ddev->dev);
>   
> @@ -2576,7 +2745,7 @@ static ssize_t amdgpu_hwmon_set_fan1_target(struct device *dev,
>   	u32 value;
>   	u32 pwm_mode;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	err = pm_runtime_get_sync(adev->ddev->dev);
> @@ -2585,11 +2754,15 @@ static ssize_t amdgpu_hwmon_set_fan1_target(struct device *dev,
>   		return err;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev))
>   		pwm_mode = smu_get_fan_control_mode(&adev->smu);
>   	else
>   		pwm_mode = amdgpu_dpm_get_fan_control_mode(adev);
>   
> +	up_read(&adev->reset_sem);
> +
>   	if (pwm_mode != AMD_FAN_CTRL_MANUAL) {
>   		pm_runtime_mark_last_busy(adev->ddev->dev);
>   		pm_runtime_put_autosuspend(adev->ddev->dev);
> @@ -2603,6 +2776,8 @@ static ssize_t amdgpu_hwmon_set_fan1_target(struct device *dev,
>   		return err;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev))
>   		err = smu_set_fan_speed_rpm(&adev->smu, value);
>   	else if (adev->powerplay.pp_funcs->set_fan_speed_rpm)
> @@ -2610,6 +2785,8 @@ static ssize_t amdgpu_hwmon_set_fan1_target(struct device *dev,
>   	else
>   		err = -EINVAL;
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(adev->ddev->dev);
>   	pm_runtime_put_autosuspend(adev->ddev->dev);
>   
> @@ -2627,7 +2804,7 @@ static ssize_t amdgpu_hwmon_get_fan1_enable(struct device *dev,
>   	u32 pwm_mode = 0;
>   	int ret;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	ret = pm_runtime_get_sync(adev->ddev->dev);
> @@ -2636,18 +2813,23 @@ static ssize_t amdgpu_hwmon_get_fan1_enable(struct device *dev,
>   		return ret;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev)) {
>   		pwm_mode = smu_get_fan_control_mode(&adev->smu);
>   	} else {
>   		if (!adev->powerplay.pp_funcs->get_fan_control_mode) {
>   			pm_runtime_mark_last_busy(adev->ddev->dev);
>   			pm_runtime_put_autosuspend(adev->ddev->dev);
> +			up_read(&adev->reset_sem);
>   			return -EINVAL;
>   		}
>   
>   		pwm_mode = amdgpu_dpm_get_fan_control_mode(adev);
>   	}
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(adev->ddev->dev);
>   	pm_runtime_put_autosuspend(adev->ddev->dev);
>   
> @@ -2664,7 +2846,7 @@ static ssize_t amdgpu_hwmon_set_fan1_enable(struct device *dev,
>   	int value;
>   	u32 pwm_mode;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	err = kstrtoint(buf, 10, &value);
> @@ -2684,17 +2866,22 @@ static ssize_t amdgpu_hwmon_set_fan1_enable(struct device *dev,
>   		return err;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev)) {
>   		smu_set_fan_control_mode(&adev->smu, pwm_mode);
>   	} else {
>   		if (!adev->powerplay.pp_funcs->set_fan_control_mode) {
>   			pm_runtime_mark_last_busy(adev->ddev->dev);
>   			pm_runtime_put_autosuspend(adev->ddev->dev);
> +			up_read(&adev->reset_sem);
>   			return -EINVAL;
>   		}
>   		amdgpu_dpm_set_fan_control_mode(adev, pwm_mode);
>   	}
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(adev->ddev->dev);
>   	pm_runtime_put_autosuspend(adev->ddev->dev);
>   
> @@ -2709,7 +2896,7 @@ static ssize_t amdgpu_hwmon_show_vddgfx(struct device *dev,
>   	u32 vddgfx;
>   	int r, size = sizeof(vddgfx);
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	r = pm_runtime_get_sync(adev->ddev->dev);
> @@ -2718,9 +2905,11 @@ static ssize_t amdgpu_hwmon_show_vddgfx(struct device *dev,
>   		return r;
>   	}
>   
> +	down_read(&adev->reset_sem);
>   	/* get the voltage */
>   	r = amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_VDDGFX,
>   				   (void *)&vddgfx, &size);
> +	up_read(&adev->reset_sem);
>   
>   	pm_runtime_mark_last_busy(adev->ddev->dev);
>   	pm_runtime_put_autosuspend(adev->ddev->dev);
> @@ -2746,7 +2935,7 @@ static ssize_t amdgpu_hwmon_show_vddnb(struct device *dev,
>   	u32 vddnb;
>   	int r, size = sizeof(vddnb);
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	/* only APUs have vddnb */
> @@ -2759,9 +2948,11 @@ static ssize_t amdgpu_hwmon_show_vddnb(struct device *dev,
>   		return r;
>   	}
>   
> +	down_read(&adev->reset_sem);
>   	/* get the voltage */
>   	r = amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_VDDNB,
>   				   (void *)&vddnb, &size);
> +	up_read(&adev->reset_sem);
>   
>   	pm_runtime_mark_last_busy(adev->ddev->dev);
>   	pm_runtime_put_autosuspend(adev->ddev->dev);
> @@ -2788,7 +2979,7 @@ static ssize_t amdgpu_hwmon_show_power_avg(struct device *dev,
>   	int r, size = sizeof(u32);
>   	unsigned uw;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	r = pm_runtime_get_sync(adev->ddev->dev);
> @@ -2797,9 +2988,11 @@ static ssize_t amdgpu_hwmon_show_power_avg(struct device *dev,
>   		return r;
>   	}
>   
> +	down_read(&adev->reset_sem);
>   	/* get the voltage */
>   	r = amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_GPU_POWER,
>   				   (void *)&query, &size);
> +	up_read(&adev->reset_sem);
>   
>   	pm_runtime_mark_last_busy(adev->ddev->dev);
>   	pm_runtime_put_autosuspend(adev->ddev->dev);
> @@ -2829,7 +3022,7 @@ static ssize_t amdgpu_hwmon_show_power_cap_max(struct device *dev,
>   	ssize_t size;
>   	int r;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	r = pm_runtime_get_sync(adev->ddev->dev);
> @@ -2838,6 +3031,8 @@ static ssize_t amdgpu_hwmon_show_power_cap_max(struct device *dev,
>   		return r;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev)) {
>   		smu_get_power_limit(&adev->smu, &limit, true);
>   		size = snprintf(buf, PAGE_SIZE, "%u\n", limit * 1000000);
> @@ -2848,6 +3043,8 @@ static ssize_t amdgpu_hwmon_show_power_cap_max(struct device *dev,
>   		size = snprintf(buf, PAGE_SIZE, "\n");
>   	}
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(adev->ddev->dev);
>   	pm_runtime_put_autosuspend(adev->ddev->dev);
>   
> @@ -2863,7 +3060,7 @@ static ssize_t amdgpu_hwmon_show_power_cap(struct device *dev,
>   	ssize_t size;
>   	int r;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	r = pm_runtime_get_sync(adev->ddev->dev);
> @@ -2872,6 +3069,8 @@ static ssize_t amdgpu_hwmon_show_power_cap(struct device *dev,
>   		return r;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev)) {
>   		smu_get_power_limit(&adev->smu, &limit, false);
>   		size = snprintf(buf, PAGE_SIZE, "%u\n", limit * 1000000);
> @@ -2882,6 +3081,8 @@ static ssize_t amdgpu_hwmon_show_power_cap(struct device *dev,
>   		size = snprintf(buf, PAGE_SIZE, "\n");
>   	}
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(adev->ddev->dev);
>   	pm_runtime_put_autosuspend(adev->ddev->dev);
>   
> @@ -2898,7 +3099,7 @@ static ssize_t amdgpu_hwmon_set_power_cap(struct device *dev,
>   	int err;
>   	u32 value;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	if (amdgpu_sriov_vf(adev))
> @@ -2917,6 +3118,8 @@ static ssize_t amdgpu_hwmon_set_power_cap(struct device *dev,
>   		return err;
>   	}
>   
> +	down_read(&adev->reset_sem);
> +
>   	if (is_support_sw_smu(adev))
>   		err = smu_set_power_limit(&adev->smu, value);
>   	else if (adev->powerplay.pp_funcs && adev->powerplay.pp_funcs->set_power_limit)
> @@ -2924,6 +3127,8 @@ static ssize_t amdgpu_hwmon_set_power_cap(struct device *dev,
>   	else
>   		err = -EINVAL;
>   
> +	up_read(&adev->reset_sem);
> +
>   	pm_runtime_mark_last_busy(adev->ddev->dev);
>   	pm_runtime_put_autosuspend(adev->ddev->dev);
>   
> @@ -2941,7 +3146,7 @@ static ssize_t amdgpu_hwmon_show_sclk(struct device *dev,
>   	uint32_t sclk;
>   	int r, size = sizeof(sclk);
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	r = pm_runtime_get_sync(adev->ddev->dev);
> @@ -2950,9 +3155,11 @@ static ssize_t amdgpu_hwmon_show_sclk(struct device *dev,
>   		return r;
>   	}
>   
> +	down_read(&adev->reset_sem);
>   	/* get the sclk */
>   	r = amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_GFX_SCLK,
>   				   (void *)&sclk, &size);
> +	up_read(&adev->reset_sem);
>   
>   	pm_runtime_mark_last_busy(adev->ddev->dev);
>   	pm_runtime_put_autosuspend(adev->ddev->dev);
> @@ -2978,7 +3185,7 @@ static ssize_t amdgpu_hwmon_show_mclk(struct device *dev,
>   	uint32_t mclk;
>   	int r, size = sizeof(mclk);
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	r = pm_runtime_get_sync(adev->ddev->dev);
> @@ -2987,9 +3194,11 @@ static ssize_t amdgpu_hwmon_show_mclk(struct device *dev,
>   		return r;
>   	}
>   
> +	down_read(&adev->reset_sem);
>   	/* get the sclk */
>   	r = amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_GFX_MCLK,
>   				   (void *)&mclk, &size);
> +	up_read(&adev->reset_sem);
>   
>   	pm_runtime_mark_last_busy(adev->ddev->dev);
>   	pm_runtime_put_autosuspend(adev->ddev->dev);
> @@ -3870,7 +4079,7 @@ static int amdgpu_debugfs_pm_info(struct seq_file *m, void *data)
>   	u32 flags = 0;
>   	int r;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EPERM;
>   
>   	r = pm_runtime_get_sync(dev->dev);
> @@ -3879,7 +4088,10 @@ static int amdgpu_debugfs_pm_info(struct seq_file *m, void *data)
>   		return r;
>   	}
>   
> +	down_read(&adev->reset_sem);
>   	amdgpu_device_ip_get_clockgating_state(adev, &flags);
> +	up_read(&adev->reset_sem);
> +
>   	seq_printf(m, "Clock Gating Flags Mask: 0x%x\n", flags);
>   	amdgpu_parse_cg_state(m, flags);
>   	seq_printf(m, "\n");
> @@ -3891,6 +4103,7 @@ static int amdgpu_debugfs_pm_info(struct seq_file *m, void *data)
>   		return 0;
>   	}
>   
> +	down_read(&adev->reset_sem);
>   	if (!is_support_sw_smu(adev) &&
>   	    adev->powerplay.pp_funcs->debugfs_print_current_performance_level) {
>   		mutex_lock(&adev->pm.mutex);
> @@ -3903,6 +4116,7 @@ static int amdgpu_debugfs_pm_info(struct seq_file *m, void *data)
>   	} else {
>   		r = amdgpu_debugfs_pm_info_pp(m, adev);
>   	}
> +	up_read(&adev->reset_sem);
>   
>   	pm_runtime_mark_last_busy(dev->dev);
>   	pm_runtime_put_autosuspend(dev->dev);
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
> index e20695b44dbe..ab0883eeadb3 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
> @@ -1682,7 +1682,7 @@ static int psp_load_smu_fw(struct psp_context *psp)
>   		return 0;
>   
>   
> -	if (adev->in_gpu_reset && ras && ras->supported) {
> +	if (atomic_read(&adev->in_gpu_reset) && ras && ras->supported) {
>   		ret = amdgpu_dpm_set_mp1_state(adev, PP_MP1_STATE_UNLOAD);
>   		if (ret) {
>   			DRM_WARN("Failed to set MP1 state prepare for reload\n");
> @@ -1796,7 +1796,7 @@ static int psp_load_fw(struct amdgpu_device *adev)
>   	int ret;
>   	struct psp_context *psp = &adev->psp;
>   
> -	if (amdgpu_sriov_vf(adev) && adev->in_gpu_reset) {
> +	if (amdgpu_sriov_vf(adev) && atomic_read(&adev->in_gpu_reset)) {
>   		psp_ring_stop(psp, PSP_RING_TYPE__KM); /* should not destroy ring, only stop */
>   		goto skip_memalloc;
>   	}
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
> index ab8e7c91c645..3e534262e381 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
> @@ -1975,7 +1975,7 @@ int amdgpu_ras_late_init(struct amdgpu_device *adev,
>   			amdgpu_ras_request_reset_on_boot(adev,
>   					ras_block->block);
>   			return 0;
> -		} else if (adev->in_suspend || adev->in_gpu_reset) {
> +		} else if (adev->in_suspend || atomic_read(&adev->in_gpu_reset)) {
>   			/* in resume phase, if fail to enable ras,
>   			 * clean up all ras fs nodes, and disable ras */
>   			goto cleanup;
> @@ -1984,7 +1984,7 @@ int amdgpu_ras_late_init(struct amdgpu_device *adev,
>   	}
>   
>   	/* in resume phase, no need to create ras fs node */
> -	if (adev->in_suspend || adev->in_gpu_reset)
> +	if (adev->in_suspend || atomic_read(&adev->in_gpu_reset))
>   		return 0;
>   
>   	if (ih_info->cb) {
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
> index 56beafbd3ab9..2e670aa68758 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
> @@ -2093,7 +2093,7 @@ void amdgpu_ttm_set_buffer_funcs_status(struct amdgpu_device *adev, bool enable)
>   	uint64_t size;
>   	int r;
>   
> -	if (!adev->mman.initialized || adev->in_gpu_reset ||
> +	if (!adev->mman.initialized || atomic_read(&adev->in_gpu_reset) ||
>   	    adev->mman.buffer_funcs_enabled == enable)
>   		return;
>   
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c
> index 744404a05fee..da6fa28b1b5f 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c
> @@ -628,7 +628,7 @@ int amdgpu_ucode_init_bo(struct amdgpu_device *adev)
>   	struct amdgpu_firmware_info *ucode = NULL;
>   
>    /* for baremetal, the ucode is allocated in gtt, so don't need to fill the bo when reset/suspend */
> -	if (!amdgpu_sriov_vf(adev) && (adev->in_gpu_reset || adev->in_suspend))
> +	if (!amdgpu_sriov_vf(adev) && (atomic_read(&adev->in_gpu_reset) || adev->in_suspend))
>   		return 0;
>   	/*
>   	 * if SMU loaded firmware, it needn't add SMC, UVD, and VCE
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
> index f826945989c7..336fa8ba2465 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
> @@ -325,9 +325,9 @@ static inline bool is_virtual_machine(void)
>   #define amdgpu_sriov_is_pp_one_vf(adev) \
>   	((adev)->virt.gim_feature & AMDGIM_FEATURE_PP_ONE_VF)
>   #define amdgpu_sriov_is_debug(adev) \
> -	((!adev->in_gpu_reset) && adev->virt.tdr_debug)
> +	((!atomic_read(&adev->in_gpu_reset)) && adev->virt.tdr_debug)
>   #define amdgpu_sriov_is_normal(adev) \
> -	((!adev->in_gpu_reset) && (!adev->virt.tdr_debug))
> +	((!atomic_read(&adev->in_gpu_reset)) && (!adev->virt.tdr_debug))
>   
>   bool amdgpu_virt_mmio_blocked(struct amdgpu_device *adev);
>   void amdgpu_virt_init_setting(struct amdgpu_device *adev);
> diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
> index a6170a346b39..b3c1e05da45e 100644
> --- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
> @@ -6102,7 +6102,7 @@ static int gfx_v10_0_gfx_init_queue(struct amdgpu_ring *ring)
>   	struct v10_gfx_mqd *mqd = ring->mqd_ptr;
>   	int mqd_idx = ring - &adev->gfx.gfx_ring[0];
>   
> -	if (!adev->in_gpu_reset && !adev->in_suspend) {
> +	if (!atomic_read(&adev->in_gpu_reset) && !adev->in_suspend) {
>   		memset((void *)mqd, 0, sizeof(*mqd));
>   		mutex_lock(&adev->srbm_mutex);
>   		nv_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
> @@ -6114,7 +6114,7 @@ static int gfx_v10_0_gfx_init_queue(struct amdgpu_ring *ring)
>   		mutex_unlock(&adev->srbm_mutex);
>   		if (adev->gfx.me.mqd_backup[mqd_idx])
>   			memcpy(adev->gfx.me.mqd_backup[mqd_idx], mqd, sizeof(*mqd));
> -	} else if (adev->in_gpu_reset) {
> +	} else if (atomic_read(&adev->in_gpu_reset)) {
>   		/* reset mqd with the backup copy */
>   		if (adev->gfx.me.mqd_backup[mqd_idx])
>   			memcpy(mqd, adev->gfx.me.mqd_backup[mqd_idx], sizeof(*mqd));
> @@ -6463,7 +6463,7 @@ static int gfx_v10_0_kiq_init_queue(struct amdgpu_ring *ring)
>   
>   	gfx_v10_0_kiq_setting(ring);
>   
> -	if (adev->in_gpu_reset) { /* for GPU_RESET case */
> +	if (atomic_read(&adev->in_gpu_reset)) { /* for GPU_RESET case */
>   		/* reset MQD to a clean status */
>   		if (adev->gfx.mec.mqd_backup[mqd_idx])
>   			memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(*mqd));
> @@ -6499,7 +6499,7 @@ static int gfx_v10_0_kcq_init_queue(struct amdgpu_ring *ring)
>   	struct v10_compute_mqd *mqd = ring->mqd_ptr;
>   	int mqd_idx = ring - &adev->gfx.compute_ring[0];
>   
> -	if (!adev->in_gpu_reset && !adev->in_suspend) {
> +	if (!atomic_read(&adev->in_gpu_reset) && !adev->in_suspend) {
>   		memset((void *)mqd, 0, sizeof(*mqd));
>   		mutex_lock(&adev->srbm_mutex);
>   		nv_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
> @@ -6509,7 +6509,7 @@ static int gfx_v10_0_kcq_init_queue(struct amdgpu_ring *ring)
>   
>   		if (adev->gfx.mec.mqd_backup[mqd_idx])
>   			memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(*mqd));
> -	} else if (adev->in_gpu_reset) { /* for GPU_RESET case */
> +	} else if (atomic_read(&adev->in_gpu_reset)) { /* for GPU_RESET case */
>   		/* reset MQD to a clean status */
>   		if (adev->gfx.mec.mqd_backup[mqd_idx])
>   			memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(*mqd));
> diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
> index 33f1c4a46ebe..c70f3bc5cea2 100644
> --- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
> @@ -4632,7 +4632,7 @@ static int gfx_v8_0_kiq_init_queue(struct amdgpu_ring *ring)
>   
>   	gfx_v8_0_kiq_setting(ring);
>   
> -	if (adev->in_gpu_reset) { /* for GPU_RESET case */
> +	if (atomic_read(&adev->in_gpu_reset)) { /* for GPU_RESET case */
>   		/* reset MQD to a clean status */
>   		if (adev->gfx.mec.mqd_backup[mqd_idx])
>   			memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct vi_mqd_allocation));
> @@ -4669,7 +4669,7 @@ static int gfx_v8_0_kcq_init_queue(struct amdgpu_ring *ring)
>   	struct vi_mqd *mqd = ring->mqd_ptr;
>   	int mqd_idx = ring - &adev->gfx.compute_ring[0];
>   
> -	if (!adev->in_gpu_reset && !adev->in_suspend) {
> +	if (!atomic_read(&adev->in_gpu_reset) && !adev->in_suspend) {
>   		memset((void *)mqd, 0, sizeof(struct vi_mqd_allocation));
>   		((struct vi_mqd_allocation *)mqd)->dynamic_cu_mask = 0xFFFFFFFF;
>   		((struct vi_mqd_allocation *)mqd)->dynamic_rb_mask = 0xFFFFFFFF;
> @@ -4681,7 +4681,7 @@ static int gfx_v8_0_kcq_init_queue(struct amdgpu_ring *ring)
>   
>   		if (adev->gfx.mec.mqd_backup[mqd_idx])
>   			memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(struct vi_mqd_allocation));
> -	} else if (adev->in_gpu_reset) { /* for GPU_RESET case */
> +	} else if (atomic_read(&adev->in_gpu_reset)) { /* for GPU_RESET case */
>   		/* reset MQD to a clean status */
>   		if (adev->gfx.mec.mqd_backup[mqd_idx])
>   			memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct vi_mqd_allocation));
> diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
> index cb9d60a4e05e..9020a0f40e28 100644
> --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
> @@ -3684,7 +3684,7 @@ static int gfx_v9_0_kiq_init_queue(struct amdgpu_ring *ring)
>   
>   	gfx_v9_0_kiq_setting(ring);
>   
> -	if (adev->in_gpu_reset) { /* for GPU_RESET case */
> +	if (atomic_read(&adev->in_gpu_reset)) { /* for GPU_RESET case */
>   		/* reset MQD to a clean status */
>   		if (adev->gfx.mec.mqd_backup[mqd_idx])
>   			memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct v9_mqd_allocation));
> @@ -3722,7 +3722,7 @@ static int gfx_v9_0_kcq_init_queue(struct amdgpu_ring *ring)
>   	struct v9_mqd *mqd = ring->mqd_ptr;
>   	int mqd_idx = ring - &adev->gfx.compute_ring[0];
>   
> -	if (!adev->in_gpu_reset && !adev->in_suspend) {
> +	if (!atomic_read(&adev->in_gpu_reset) && !adev->in_suspend) {
>   		memset((void *)mqd, 0, sizeof(struct v9_mqd_allocation));
>   		((struct v9_mqd_allocation *)mqd)->dynamic_cu_mask = 0xFFFFFFFF;
>   		((struct v9_mqd_allocation *)mqd)->dynamic_rb_mask = 0xFFFFFFFF;
> @@ -3734,7 +3734,7 @@ static int gfx_v9_0_kcq_init_queue(struct amdgpu_ring *ring)
>   
>   		if (adev->gfx.mec.mqd_backup[mqd_idx])
>   			memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(struct v9_mqd_allocation));
> -	} else if (adev->in_gpu_reset) { /* for GPU_RESET case */
> +	} else if (atomic_read(&adev->in_gpu_reset)) { /* for GPU_RESET case */
>   		/* reset MQD to a clean status */
>   		if (adev->gfx.mec.mqd_backup[mqd_idx])
>   			memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct v9_mqd_allocation));
> @@ -3928,7 +3928,7 @@ static int gfx_v9_0_hw_fini(void *handle)
>   	/* Use deinitialize sequence from CAIL when unbinding device from driver,
>   	 * otherwise KIQ is hanging when binding back
>   	 */
> -	if (!adev->in_gpu_reset && !adev->in_suspend) {
> +	if (!atomic_read(&adev->in_gpu_reset) && !adev->in_suspend) {
>   		mutex_lock(&adev->srbm_mutex);
>   		soc15_grbm_select(adev, adev->gfx.kiq.ring.me,
>   				adev->gfx.kiq.ring.pipe,
> @@ -4086,7 +4086,7 @@ static uint64_t gfx_v9_0_kiq_read_clock(struct amdgpu_device *adev)
>   	 *
>   	 * also don't wait anymore for IRQ context
>   	 * */
> -	if (r < 1 && (adev->in_gpu_reset || in_interrupt()))
> +	if (r < 1 && (atomic_read(&adev->in_gpu_reset) || in_interrupt()))
>   		goto failed_kiq_read;
>   
>   	might_sleep();
> diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
> index 8f29f79f820d..b761bf59ed11 100644
> --- a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
> @@ -356,7 +356,7 @@ static void gmc_v10_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
>   	 */
>   	if (adev->gfx.kiq.ring.sched.ready &&
>   	    (amdgpu_sriov_runtime(adev) || !amdgpu_sriov_vf(adev)) &&
> -	    !adev->in_gpu_reset) {
> +	    !atomic_read(&adev->in_gpu_reset)) {
>   
>   		struct amdgpu_vmhub *hub = &adev->vmhub[vmhub];
>   		const unsigned eng = 17;
> @@ -381,7 +381,7 @@ static void gmc_v10_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
>   
>   	if (!adev->mman.buffer_funcs_enabled ||
>   	    !adev->ib_pool_ready ||
> -	    adev->in_gpu_reset ||
> +	    atomic_read(&adev->in_gpu_reset) ||
>   	    ring->sched.ready == false) {
>   		gmc_v10_0_flush_vm_hub(adev, vmid, AMDGPU_GFXHUB_0, 0);
>   		mutex_unlock(&adev->mman.gtt_window_lock);
> diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
> index e18296dc1386..d83b83459ba7 100644
> --- a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
> @@ -434,7 +434,7 @@ static int gmc_v7_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev,
>   	int vmid;
>   	unsigned int tmp;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EIO;
>   
>   	for (vmid = 1; vmid < 16; vmid++) {
> diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
> index a9e722b8a458..5854df6dcc32 100644
> --- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
> @@ -635,7 +635,7 @@ static int gmc_v8_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev,
>   	int vmid;
>   	unsigned int tmp;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EIO;
>   
>   	for (vmid = 1; vmid < 16; vmid++) {
> diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
> index 6e4f3ff4810f..b82a1af86f43 100644
> --- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
> @@ -501,7 +501,7 @@ static void gmc_v9_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
>   	 */
>   	if (adev->gfx.kiq.ring.sched.ready &&
>   			(amdgpu_sriov_runtime(adev) || !amdgpu_sriov_vf(adev)) &&
> -			!adev->in_gpu_reset) {
> +			!atomic_read(&adev->in_gpu_reset)) {
>   		uint32_t req = hub->vm_inv_eng0_req + hub->eng_distance * eng;
>   		uint32_t ack = hub->vm_inv_eng0_ack + hub->eng_distance * eng;
>   
> @@ -596,7 +596,7 @@ static int gmc_v9_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev,
>   	struct amdgpu_ring *ring = &adev->gfx.kiq.ring;
>   	struct amdgpu_kiq *kiq = &adev->gfx.kiq;
>   
> -	if (adev->in_gpu_reset)
> +	if (atomic_read(&adev->in_gpu_reset))
>   		return -EIO;
>   
>   	if (ring->sched.ready) {
> diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
> index 5fd67e1cc2a0..fe31cbeccfe9 100644
> --- a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
> +++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
> @@ -238,20 +238,16 @@ static void xgpu_ai_mailbox_flr_work(struct work_struct *work)
>   	struct amdgpu_virt *virt = container_of(work, struct amdgpu_virt, flr_work);
>   	struct amdgpu_device *adev = container_of(virt, struct amdgpu_device, virt);
>   	int timeout = AI_MAILBOX_POLL_FLR_TIMEDOUT;
> -	int locked;
>   
>   	/* block amdgpu_gpu_recover till msg FLR COMPLETE received,
>   	 * otherwise the mailbox msg will be ruined/reseted by
>   	 * the VF FLR.
>   	 *
> -	 * we can unlock the lock_reset to allow "amdgpu_job_timedout"
> +	 * we can unlock the reset_sem to allow "amdgpu_job_timedout"
>   	 * to run gpu_recover() after FLR_NOTIFICATION_CMPL received
>   	 * which means host side had finished this VF's FLR.
>   	 */
> -	locked = mutex_trylock(&adev->lock_reset);
> -	if (locked)
> -		adev->in_gpu_reset = true;
> -
> +	down_read(&adev->reset_sem);
>   	do {
>   		if (xgpu_ai_mailbox_peek_msg(adev) == IDH_FLR_NOTIFICATION_CMPL)
>   			goto flr_done;
> @@ -261,10 +257,7 @@ static void xgpu_ai_mailbox_flr_work(struct work_struct *work)
>   	} while (timeout > 1);
>   
>   flr_done:
> -	if (locked) {
> -		adev->in_gpu_reset = false;
> -		mutex_unlock(&adev->lock_reset);
> -	}
> +	up_read(&adev->reset_sem);
>   
>   	/* Trigger recovery for world switch failure if no TDR */
>   	if (amdgpu_device_should_recover_gpu(adev)
> diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c
> index ce2bf1fb79ed..6f55172e8337 100644
> --- a/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c
> +++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c
> @@ -259,20 +259,16 @@ static void xgpu_nv_mailbox_flr_work(struct work_struct *work)
>   	struct amdgpu_virt *virt = container_of(work, struct amdgpu_virt, flr_work);
>   	struct amdgpu_device *adev = container_of(virt, struct amdgpu_device, virt);
>   	int timeout = NV_MAILBOX_POLL_FLR_TIMEDOUT;
> -	int locked;
>   
>   	/* block amdgpu_gpu_recover till msg FLR COMPLETE received,
>   	 * otherwise the mailbox msg will be ruined/reseted by
>   	 * the VF FLR.
>   	 *
> -	 * we can unlock the lock_reset to allow "amdgpu_job_timedout"
> +	 * we can unlock the reset_sem to allow "amdgpu_job_timedout"
>   	 * to run gpu_recover() after FLR_NOTIFICATION_CMPL received
>   	 * which means host side had finished this VF's FLR.
>   	 */
> -	locked = mutex_trylock(&adev->lock_reset);
> -	if (locked)
> -		adev->in_gpu_reset = true;
> -
> +	down_read(&adev->reset_sem);
>   	do {
>   		if (xgpu_nv_mailbox_peek_msg(adev) == IDH_FLR_NOTIFICATION_CMPL)
>   			goto flr_done;
> @@ -282,10 +278,7 @@ static void xgpu_nv_mailbox_flr_work(struct work_struct *work)
>   	} while (timeout > 1);
>   
>   flr_done:
> -	if (locked) {
> -		adev->in_gpu_reset = false;
> -		mutex_unlock(&adev->lock_reset);
> -	}
> +	up_read(&adev->reset_sem);
>   
>   	/* Trigger recovery for world switch failure if no TDR */
>   	if (amdgpu_device_should_recover_gpu(adev)
> diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
> index e9b96ad3d9a5..393642c32387 100644
> --- a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
> +++ b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
> @@ -1482,13 +1482,20 @@ static int kfd_ioctl_map_memory_to_gpu(struct file *filep,
>   
>   	/* Flush TLBs after waiting for the page table updates to complete */
>   	for (i = 0; i < args->n_devices; i++) {
> +		struct device_queue_manager *dqm;
> +
>   		peer = kfd_device_by_id(devices_arr[i]);
>   		if (WARN_ON_ONCE(!peer))
>   			continue;
>   		peer_pdd = kfd_get_process_device_data(peer, p);
>   		if (WARN_ON_ONCE(!peer_pdd))
>   			continue;
> +
> +		dqm = peer_pdd->dev->dqm;
> +
> +		dqm_lock(dqm);
>   		kfd_flush_tlb(peer_pdd);
> +		dqm_unlock(dqm);
>   	}
>   
>   	kfree(devices_arr);
> diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
> index dd550025d1c1..0b5409cede52 100644
> --- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
> +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
> @@ -304,15 +304,17 @@ static void deallocate_vmid(struct device_queue_manager *dqm,
>   				struct qcm_process_device *qpd,
>   				struct queue *q)
>   {
> -	/* On GFX v7, CP doesn't flush TC at dequeue */
> -	if (q->device->device_info->asic_family == CHIP_HAWAII)
> -		if (flush_texture_cache_nocpsch(q->device, qpd))
> -			pr_err("Failed to flush TC\n");
> +	if (!dqm->is_resetting) {
> +		/* On GFX v7, CP doesn't flush TC at dequeue */
> +		if (q->device->device_info->asic_family == CHIP_HAWAII)
> +			if (flush_texture_cache_nocpsch(q->device, qpd))
> +				pr_err("Failed to flush TC\n");
>   
> -	kfd_flush_tlb(qpd_to_pdd(qpd));
> +		kfd_flush_tlb(qpd_to_pdd(qpd));
>   
> -	/* Release the vmid mapping */
> -	set_pasid_vmid_mapping(dqm, 0, qpd->vmid);
> +		/* Release the vmid mapping */
> +		set_pasid_vmid_mapping(dqm, 0, qpd->vmid);
> +	}
>   	dqm->vmid_pasid[qpd->vmid] = 0;
>   
>   	qpd->vmid = 0;
> diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process.c b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
> index 013c2b018edc..71be897d4c2a 100644
> --- a/drivers/gpu/drm/amd/amdkfd/kfd_process.c
> +++ b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
> @@ -1551,6 +1551,10 @@ int kfd_reserved_mem_mmap(struct kfd_dev *dev, struct kfd_process *process,
>   void kfd_flush_tlb(struct kfd_process_device *pdd)
>   {
>   	struct kfd_dev *dev = pdd->dev;
> +	struct device_queue_manager *dqm = dev->dqm;
> +
> +	if (dqm->is_resetting)
> +		return;
>   
>   	if (dev->dqm->sched_policy == KFD_SCHED_POLICY_NO_HWS) {
>   		/* Nothing to flush until a VMID is assigned, which
> diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
> index b4e120e95438..db332282c01e 100644
> --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
> +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
> @@ -1640,7 +1640,7 @@ static int dm_suspend(void *handle)
>   	struct amdgpu_display_manager *dm = &adev->dm;
>   	int ret = 0;
>   
> -	if (adev->in_gpu_reset) {
> +	if (atomic_read(&adev->in_gpu_reset)) {
>   		mutex_lock(&dm->dc_lock);
>   		dm->cached_dc_state = dc_copy_state(dm->dc->current_state);
>   
> @@ -1826,7 +1826,7 @@ static int dm_resume(void *handle)
>   	struct dc_state *dc_state;
>   	int i, r, j;
>   
> -	if (adev->in_gpu_reset) {
> +	if (atomic_read(&adev->in_gpu_reset)) {
>   		dc_state = dm->cached_dc_state;
>   
>   		r = dm_dmub_hw_init(adev);
> diff --git a/drivers/gpu/drm/amd/powerplay/amdgpu_smu.c b/drivers/gpu/drm/amd/powerplay/amdgpu_smu.c
> index fe4948aa662f..69134a0d660c 100644
> --- a/drivers/gpu/drm/amd/powerplay/amdgpu_smu.c
> +++ b/drivers/gpu/drm/amd/powerplay/amdgpu_smu.c
> @@ -1341,7 +1341,7 @@ static int smu_disable_dpms(struct smu_context *smu)
>   	uint64_t features_to_disable;
>   	int ret = 0;
>   	bool use_baco = !smu->is_apu &&
> -		((adev->in_gpu_reset &&
> +		((atomic_read(&adev->in_gpu_reset) &&
>   		  (amdgpu_asic_reset_method(adev) == AMD_RESET_METHOD_BACO)) ||
>   		 ((adev->in_runpm || adev->in_hibernate) && amdgpu_asic_supports_baco(adev)));
>   
> diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_hwmgr.c
> index 3b8839641770..13e304066d94 100644
> --- a/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_hwmgr.c
> +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_hwmgr.c
> @@ -484,7 +484,7 @@ static int vega20_setup_asic_task(struct pp_hwmgr *hwmgr)
>   {
>   	struct amdgpu_device *adev = (struct amdgpu_device *)(hwmgr->adev);
>   	int ret = 0;
> -	bool use_baco = (adev->in_gpu_reset &&
> +	bool use_baco = (atomic_read(&adev->in_gpu_reset) &&
>   			 (amdgpu_asic_reset_method(adev) == AMD_RESET_METHOD_BACO)) ||
>   		(adev->in_runpm && amdgpu_asic_supports_baco(adev));
>   
_______________________________________________
amd-gfx mailing list
amd-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/amd-gfx